社交网络环境下用户个性化需求的服务发现与聚合方法研究

A rapid development of social networking has given rise to a new e-commerce paradigm called social commerce. Social commerce has emerged as the major model of e-commerce today. However, most of social commerce provide generalized services and business processing for all service objects, lacking of characteristics and pertinence, and are hard to meet users’ diverse needs of information and services. A study of social commerce personalized service discovery and service aggregation algorithm has become one of the most challenging and difficult study within this field. In response to these challenges, this project aims to study personalized service discovery and service aggregation algorithm in social commerce. This project firstly establishes an individual differences classification model by using the probabilistic clustering and the Partition-based Clustering methods, which can build the foundation for acquiring user personalized needs. To support service accurate classification and description in service discovery as well as recommendation systems, a personalized service model is built based on a user-driven approach and the personalized service ontology initialization. Then, a three-level of service similarity computing method is developed and used, including service name, service attributes and service relationship, with the purpose of achieving a personalized service discovery method that can meet users’ needs. The results of this project will improve the accurate positioning of service targets in service resources and its efficiency, providing theoretical basis for the realization of personalized and intelligent social commerce services, which has the important theoretical significance and wide application prospect.

社交网络的迅速发展与广泛应用，使电子商务呈现出全新的商务模式，即：社交商务。社交商务已成为目前电子商务的主要模式。然而，大多数社交商务服务提供大众化、面向所有服务对象的通用服务和业务处理，缺乏特色和针对性，无法满足用户特定的信息与服务需求。个性化服务发现与聚合方法，已成为该领域研究的热点和难点之一。本课题针对这一难点，拟采用概率聚类和分区聚类挖掘方法，建立用户个性化差异分类模型，为获取用户个性化需求建立基础；以用户驱动和个性化服务本体生成构建服务模型，为服务发现与推荐系统中服务的准确分类和描述提供支持；利用服务名称、属性及关联关系的三级服务相似度计算方法，为服务描述与用户服务请求的相关性判断提供判据。该研究成果，将提高服务资源中服务目标的精准定位，实现服务资源的高效利用，为社交网络的服务个性化与智能化不断融合提供新的理论参考。

本课题研究针对社交网络环境下商务服务活动缺乏特色和针对性，服务资源与用户特定需求的不匹配问题，研究面向用户个性化需求的服务发现与聚合方法问题，重点突破复杂社交网络环境下的个性化与智能化服务相融合的服务发现与聚合关键技术。具体包括研究了社交商务系统中用户偏好、行为、认知趋向规律，设计了用户个性化差异分类模型、行为模型和兴趣度模型，实现了面向用户个性化需求的用户模型构建；根据服务发现与流程描述，提出了社交网络服务本体框架设计，构建了社交网络环境下个性化服务本体模型，为服务本体框架提供个性化服务本体生成与优化方法；根据用户个性化需求的形式化表述，结合用户偏好与情境，建立了基于服务名称、服务属性和服务关系的三级服务相似度计算方法与基于逻辑维度和资源维度相的服务聚合方法。从而提高在社交网络服务资源中对所需服务目标的精准定位与呈现，实现服务资源的高效利用。基于本项目部分研究成果，现已发表论文24篇，其中SCI论文9篇，SSCI论文2篇，EI论文13篇; 出版学术专著1部; 授权国家发明专利2项; 完成该领域2名学术型硕士研究生培养。